I would think the spurious disc (not the Airy disc) would be the smallest resolvable feature possible, yet we can see objects 4x smaller than the Airy disc.
To begin with, the spurious disc is already about half the diameter, so you're halfway there, already. In point source resolution without a black space, there is plenty of empirical evidence suggesting very tight double stars can be 'resolved' down to 0.5 times the Raleigh limit - 1/4th the Airy disc diameter. (My personal best is 0.62 times Raleigh limit no doubt influenced by the presence of an obstruction, as is the case with 7 Tau easily split with sufficient dark space to indicate something smaller than Dawes is possible.)
At this spacial frequency, there is not sufficient contrast between to brightly lit spurious discs to show a dark space for hard resolution. And we're well beyond Dawes at this point, anyway. This is a different animal than an extended object, but the example shows the Airy disc is not the limiting feature we think it is.
As an extended object begins to exceed 1/4th the Airy disc diameter (the working definition of an optical point source), it's PSF begins to expand noticeably. When the disc radius is equal to Lambda/F it's FWHM (which is an approximation of what we see near Dawes) is much larger than the PSF of the point source Airy disc at FWHM by a factor of two. In other words, Ganymede is twice the diameter of the spurious disc leaving plenty of room for a high (enough) contrast feature to offer an Airy disc of it's own superimposed on the twice as large FWHM of Ganymede.
It's in the form of a gradient, as Brian says. The high contrast bright feature can peak above the surface intensity of Ganymede, as I understand it. And if the peak is high enough (contrast), such a bright object can be seen on the expanded PSF on an object of Airy disc diameter (whose PSF is twice FWHM of a point source Airy disc.) A dark object, too, if it's of sufficient contrast. Otherwise it might appear as an intensity fall off on one hemisphere if the feature is large enough. Galileo Regio is pretty large and seen as a less bright hemisphere (or limb shading), Perrine Regio is not large and was not seen.
If you have ever seen the diffraction rings around Jovian moons, you will note Io and Europa are more star-like in appearance. Ganymede's rings are more washed out indicating it is, indeed, not a point source and therefore does not offer a point source PSF. Io and Europa are at a diameter that is roughly 1" arc, or about half the Airy disc diameter, and their PSF is barely enlarged beyond that of a point source.
One might even resolve a very high contrast feature on either Io or Europa if one existed on their surface. The resolution would be very difficult and very similar to a very tight equally bright double with a separation near half the Raleigh limit. Io does appear elongated and this may be the cause. Resolution on Io? Surely I jest (
) I dunno, maybe. Depends on the definition of resolution. Maybe from the behavior of its PSF we can say we resolved it's brighter equator from its darker poles even though we cannot see what's actually going on. Nothing is actually 'split', it's simply elongated.
That's theory and accords with my experience with two bright crater 'specks' seen on Ganymede's surface when seeing is at least diffraction limited. Any induced aberration makes detection that much more difficult.
This is high resolution applicable to lunar observing where contrasts are very high. You are correct, the obstruction makes a tiny difference of about 10% in the realm of the very tiny (near the Airy disc and inside the first ring.) That is a difference between resolving a crater that subtends 1" arc and one that subtends 0.9" arc (which turns out to be a the difference between crater about 1.1 mile in diameter and one that is 1 mile in diameter at the lunar mean distance.) It minor, but doable when your scope is operating in near lab like conditions in the real world. I have seen both "e" on Plato, and IIRC, one closer to 1 mile elsewhere, and surface high (enough) contrast features on Ganymede.Edit: I was upwards of 400x on Ganymede. Of course, there is no further resolution to be had, only image scale and brightness as you say. But, the relative contrast should remain unchanged and the dimming does have (unknown to me) physiological effects. Being up that high didn't seem to hurt anything, it was just easier to look at. I'd assert the presence of an obstruction was helpful, if minor. It was minor enough.
Anyway, it's a great discussion and I think it applies directly to the high resolution needed for small Plato craterlets.